feat(mcp): the bundle's map, and a working method that reads it first

C5. `bundlemap.build_map` lists a bundle in its own words: one line per
source document -- its name, then the titles of its concepts in document
order -- and documents whose names differ only in their numbers (a changelog
per release, a note per week) as ONE line: the name with every number as `#`,
the count, the first and last by natural order, and the titles across the
series that are words. `SERIES_MIN` = 5, at most `TITLES_PER_LINE` = 24 titles
a line, the lines capped at `MAP_MAX_BYTES` = 48 000 together with
`lines_truncated` counting the rest. Derived on every call, never stored.

The card (`okf card`, `okf_describe`) carries it as `map` and no longer
carries `source_files`: that list named every document a second time with no
series collapsed, a quarter of the reply on a large bundle, for names the map
already carries. Chose removal over keeping both because the describe reply
has to fit a client's tool-reply limit and the map says more.

The working method now reads: take the map first (`okf card`, or
`okf_describe`), write two to four sub-questions in its words, and send them
in ONE call (`--question` repeated, or `okf_ask` `questions`). Changed in the
skill template, the generated `skills/okf-consume`, and the server
instructions (held under the 2 KB a client keeps). The regeneration recipe for
`skills/okf-consume` gains `--for-bundle`: since v1.1 the generator writes the
generic skill by default, so the recipe as published produced the other file.

A test holds a four-sub-question `okf_ask` over concepts far over the passage
size under 50 000 bytes of reply text (25 000 tokens at a pessimistic two
bytes a token). The real-collection measurements are kept in local state.

Suite on a clean tree after `git add`: 2429 passed, 2 skipped, 4 xfailed.
ruff, ruff format, mypy --strict clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Kjell Tore Guttormsen 2026-09-21 08:37:12 +02:00
commit da6faf8776
9 changed files with 450 additions and 64 deletions

View file

@ -57,28 +57,31 @@ and the cut; it decides nothing about the question.
Five steps, in this order. The pre-pass is step three, not step one: a question
asked in the wrong words reaches the wrong concepts however good the ranking is.
**1. Understand the question first.** Read the bundle's map before you search
it — what it holds, how many concepts, what the documents are called. Then put
the question into the bundle's own words: a bundle written in one language and
a question asked in another share few tokens, and the pre-pass matches tokens.
Take the terms from the bundle's own titles, not from your vocabulary.
**1. Understand the question first.** Read the bundle's `map` before you search
it — `okf card <BUNDLE_ROOT>` prints it: one line per document with its section
titles, a series of like-named documents as one line. Then put the question
into the bundle's own words: a bundle written in one language and a question
asked in another share few tokens, and the pre-pass matches tokens. Take the
terms from the map's titles, not from your vocabulary.
**2. Split a broad question into 2–4 sub-questions.** One search for a question
with four parts returns the best eight concepts for the average of the four,
which is often the best eight for none of them. Write the sub-questions down;
they are also the shape of the answer.
**3. Search per sub-question. Several searches are normal, and searching again
is allowed and expected.** Run the pre-pass once per sub-question:
**3. Search all sub-questions in ONE run. Several searches are normal, and
searching again is allowed and expected.** Give the pre-pass every sub-question
at once:
```sh
okf consume <BUNDLE_ROOT> --question "one sub-question" --out /tmp/p1.json
okf consume <BUNDLE_ROOT> --question "first sub-question" --question "second sub-question" --out /tmp/p1.json
```
After each run, read two things: what came back, and what lay just outside the
cut. `withheld.nearest` names the best-ranked concepts that missed, with their
titles — if one of them is what you were after, that is a signal about the
WORDS, not a closed door. Search again with the words that concept uses, or
Each sub-question is ranked alone and the answers are interleaved; every
excerpt names the `subquestions` it answered. After each run, read two things:
what came back, and what lay just outside the cut. `withheld.nearest` names
the best-ranked concepts that missed, with their titles — if one of them is
what you were after, that is a signal about the WORDS, not a closed door. Search again with the words that concept uses, or
ask for it by name. There is no limit on runs and no penalty for a run that
found nothing; a run that found nothing is a measurement, and its denominator
is worth carrying. When `coverage.weak` is true — a word of yours the bundle

View file

@ -68,28 +68,31 @@ and the cut; it decides nothing about the question.
Five steps, in this order. The pre-pass is step three, not step one: a question
asked in the wrong words reaches the wrong concepts however good the ranking is.
**1. Understand the question first.** Read the bundle's map before you search
it — what it holds, how many concepts, what the documents are called. Then put
the question into the bundle's own words: a bundle written in one language and
a question asked in another share few tokens, and the pre-pass matches tokens.
Take the terms from the bundle's own titles, not from your vocabulary.
**1. Understand the question first.** Read the bundle's `map` before you search
it — `okf card examples/ingest-golden-segmented-okf-v0-2/expected-bundle` prints it: one line per document with its section
titles, a series of like-named documents as one line. Then put the question
into the bundle's own words: a bundle written in one language and a question
asked in another share few tokens, and the pre-pass matches tokens. Take the
terms from the map's titles, not from your vocabulary.
**2. Split a broad question into 2–4 sub-questions.** One search for a question
with four parts returns the best eight concepts for the average of the four,
which is often the best eight for none of them. Write the sub-questions down;
they are also the shape of the answer.
**3. Search per sub-question. Several searches are normal, and searching again
is allowed and expected.** Run the pre-pass once per sub-question:
**3. Search all sub-questions in ONE run. Several searches are normal, and
searching again is allowed and expected.** Give the pre-pass every sub-question
at once:
```sh
okf consume examples/ingest-golden-segmented-okf-v0-2/expected-bundle --question "one sub-question" --out /tmp/p1.json
okf consume examples/ingest-golden-segmented-okf-v0-2/expected-bundle --question "first sub-question" --question "second sub-question" --out /tmp/p1.json
```
After each run, read two things: what came back, and what lay just outside the
cut. `withheld.nearest` names the best-ranked concepts that missed, with their
titles — if one of them is what you were after, that is a signal about the
WORDS, not a closed door. Search again with the words that concept uses, or
Each sub-question is ranked alone and the answers are interleaved; every
excerpt names the `subquestions` it answered. After each run, read two things:
what came back, and what lay just outside the cut. `withheld.nearest` names
the best-ranked concepts that missed, with their titles — if one of them is
what you were after, that is a signal about the WORDS, not a closed door. Search again with the words that concept uses, or
ask for it by name. There is no limit on runs and no penalty for a run that
found nothing; a run that found nothing is a measurement, and its denominator
is worth carrying. When `coverage.weak` is true — a word of yours the bundle

View file

@ -12,7 +12,8 @@ Regenerate them from the repository root rather than editing either file, with
```sh
okf skill examples/ingest-golden-segmented-okf-v0-2/expected-bundle \
--out skills/okf-consume --force --example-question "Hva sier veiledningen om krav?"
--out skills/okf-consume --force --for-bundle \
--example-question "Hva sier veiledningen om krav?"
python3 -c 'import os, pathlib; p = pathlib.Path("skills/okf-consume/SKILL.md"); p.write_text(p.read_text(encoding="utf-8").replace(os.path.realpath(".") + "/", ""), encoding="utf-8")'
okf check --skill skills/okf-consume/SKILL.md \
--payload skills/okf-consume/references/example-payload.json
@ -20,6 +21,8 @@ okf check --skill skills/okf-consume/SKILL.md \
Why each part is there:
- **`--for-bundle`**: since v1.1 the generator writes the GENERIC skill by
default; this copy is the instantiated one, for this bundle.
- **`--force`**: the generator refuses to replace an existing `SKILL.md`
(`refused (target_occupied)`), because a silent overwrite would destroy a
hand-edited copy.